Vegetables, Fruits, and Carotenoids The protective effect of vegetables and fruits is frequently touted as the most persuasive finding to emerge from epidemiologic studies of diet and cancer, with evidence strongest for lung and colorectal cancer. However, we recently reported that, in two large U.S. cohorts, total vegetable and fruit intake was associated with a nonsignificantly and modestly lower risk of lung cancer among women [smoking-adjusted relative risk (RR) = 0.79] and no reduction of lung cancer risk among men. Inclusion of detailed smoking history variables had substantially attenuated both protective effects. Nonetheless, in both men and women, total vegetable and fruit intake was associated with lower risk among never-smokers (RR = 0.63). In a nested case-control study of lung cancer in a cohort of Hawaiian Japanese men, individual carotenoids were measured in prediagnostic sera. Low serum levels of beta-cryptoxanthin, lycopene, and alpha-carotene, but not beta-carotene, were each modestly associated with elevated lung cancer risk (smoking-adjusted RRs = 1.3-1.5). There was no evidence of combined or synergistic effects for individual carotenoids. Thus, carotenoids, at physiologic levels, may not contribute substantially to lung cancer prevention. In the Prostate, Lung, Colorectal, and Ovarian Cancer Screening Trial (PLCO) cohort, we have now started to investigate the relationship of vegetable and fruit intake, with quantity and variety assessed in several ways, and individual carotenoids, measured in both diet and plasma, to risk of colorectal adenoma at various stages in the speculated progression to colorectal cancer. Strengths include a targeted dietary assessment instrument, substantial numbers of outcomes identified by systematic screening, and information on histology. Like other carotenoids, lycopene is a potent antioxidant. However, its structure, isomeric variability, and dietary sources are distinct. Epidemiologic research has suggested that lycopene, but not other individual carotenoids, may be associated with reduced risk of prostate cancer. In a case-control study among U.S. Blacks and Whites, serum lycopene was inversely associated with prostate cancer risk (RR~0.65), particularly so for aggressive disease (RR~0.4), in both races. This study is the first to extend this relationship to include Blacks. In study controls, serum lycopene concentrations were significantly lower (by ~18%) in Blacks than in Whites, a difference we confirmed with data from the nationally representative third National Health and Nutrition Examination Survey. Differential lycopene exposure may contribute to the racial disparity in prostate cancer incidence in the U.S. We are continuing to explore the role of lycopene in prostate carcinogenesis using high performance liquid chromatography techniques to measure lycopene geometric isomers and lycopene metabolites. Folate and One-Carbon Metabolism Folate is essential for one-carbon (methyl group) metabolism, a biochemical pathway involved in DNA synthesis, repair, and methylation. Efficient one-carbon metabolism also requires vitamins B-6 and B-12, riboflavin, and optimal activity of 10-20 enzymes. In a community-based, case-control study of invasive cervical cancer, we showed that high serum homocysteine was associated with a statistically significant 100-200% increase in risk and low serum or red blood cell folate, with only a 20-60% increase. This pattern suggests that circulating homocysteine may be an integratory measure of insufficient folate in tissues or a biomarker of disruption of one-carbon metabolism. Variant forms of two common polymorphisms in the methylene tetrahydrofolate reductase gene and a common polymorphism in the methionine synthase gene were each associated with elevated cervical cancer risk. Risk generally increased as copies of the variant gene increased. These results suggest that both genetic variability and micronutrient inadequacy in the one-carbon metabolism pathway can contribute to increased risk of cervical cancer. Additional polymorphisms in pathway genes are now being assayed. We are also exploring the role of one-carbon metabolism in the etiology of both colorectal and breast cancer in the PLCO cohort. The large number of colorectal adenomas identified will allow us to systematically search for main effects of polymorphic variation in key one-carbon metabolism genes, using spaced polymorphisms as biomarkers of genetic change. Relationships of circulating levels of homocysteine and various folate forms with genotype will also be explored. Breast Cancer and Prostate Cancer in Asian-American Populations International variation in breast cancer incidence and migrant studies indicate that modifiable factors play a major role in breast cancer etiology although the specific lifestyles and environmental exposures remain elusive. We designed a large, population-based case-control study of breast cancer in Asian-American women to take advantage of their diversity in lifestyle and breast cancer risk. Childhood, adolescent, and adult exposures were assessed by interviewing both study participants and their mothers. A six-fold gradient in breast cancer incidence by migration patterns was observed, comparable to the international differences in breast cancer rates. With the biochemical data collected, we are now focusing on relationships with migration patterns in the controls. Estrogens did not differ significantly between Asian-American women born in Asia and the West. However, androgen levels were higher among Asian-Americans born in Asia. Further efforts to understand the hormonal mechanisms underlying breast carcinogenesis should consider androgens as well as estrogens. The ratio of 2- to 16alpha-hydroxyestrone, a biomarker of estrogen metabolism, was consistently lower (by ~20%) in Asian-American women born in the West. The 2:16alpha ratio may reflect Asian lifestyles that influence estrogen metabolism and reduce breast cancer risk. Adiposity and weight gain in the decade preceding diagnosis, as well as height, were critical determinants of breast cancer risk in these Asian-American women. Thus, excess weight may function as a late stage promoter in breast carcinogenesis, and weight maintenance or reduction as an adult may have a significant and rapid impact on breast cancer risk. We are seeking biologic explanations, with emphasis initially on insulin-like growth factors (IGFs). We previously published that breast cancer risk decreased with increasing frequency of soy intake, with a statistically significant 15% reduction in risk with each additional serving per week. Continued analysis of the dietary information collected will clarify whether any other Asian or Western food groups or dietary patterns are as predictive of risk as soy. Preliminary results suggest that the relative risk of breast cancer for a positive family history is similar in Asian-American women at different levels of Westernization. This constancy implies that the lifestyles responsible for lower risk among Asians may also modify genetically determined breast cancer risk. Finally, we are now evaluating how much of the risk gradient in this migrant population is explained by parity and lactation, other menstrual and reproductive patterns, anthropometry, and unidentified factors. Foci of prostate cancer cells, like foci of breast cancer cells, seem to advance more rapidly in Western societies than in Asian societies. Elevated IGF-I has been postulated as a biomarker of more rapid progression, and possibly a cause. With prospectively stored serum samples from a cohort of Hawaiian Japanese men, we are exploring the relationship of IGF-I, IGFBP-3, free IGF-I, PSA, and proteomic patterns to risk of prostate cancer. Two important questions are whether IGF-I is equally predictive of indolent and aggressive prostate cancer and whether it is a true risk factor, preceding diagnosis by many years, or simply an early marker of preclinical disease Methods for Biochemical Epidemiology Biochemical assays can assess exposure, integrate events, and measure metabolic, and possibly causal, intermediates in epidemiologic studies. We have completed and published a series of manuscripts on the reproducibility of the kits currently used to measure a variety of estrogens, estrogen metabolites, and androgens in blood and urine from premenopausal women, postmenopausal women, and men. While some of the assays are sufficiently reliable to discriminate among individuals, others are more problematic. Comparable work on IGF-I assays indicates that absolute values can vary substantially between kits and among laboratories. With collaborators in the Center for Cancer Research, we are developing a state-of-the-art liquid chromatography/mass spectroscopy system that can measure a full range of estrogens and androgens, ~15 metabolites, and ~10 phytoestrogens in a single sample of urine. We anticipate that the approach can be extended to blood and tissue.